1 %{/* Bison Grammar Parser                             -*- C -*-
 
   3    Copyright (C) 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009 Free Software
 
   6    This file is part of Bison, the GNU Compiler Compiler.
 
   8    This program is free software: you can redistribute it and/or modify
 
   9    it under the terms of the GNU General Public License as published by
 
  10    the Free Software Foundation, either version 3 of the License, or
 
  11    (at your option) any later version.
 
  13    This program is distributed in the hope that it will be useful,
 
  14    but WITHOUT ANY WARRANTY; without even the implied warranty of
 
  15    MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 
  16    GNU General Public License for more details.
 
  18    You should have received a copy of the GNU General Public License
 
  19    along with this program.  If not, see <http://www.gnu.org/licenses/>.  */
 
  25 #include "conflicts.h"
 
  29 #include "muscle-tab.h"
 
  33 #include "scan-gram.h"
 
  34 #include "scan-code.h"
 
  36 #define YYLLOC_DEFAULT(Current, Rhs, N)  (Current) = lloc_default (Rhs, N)
 
  37 static YYLTYPE lloc_default (YYLTYPE const *, int);
 
  39 #define YY_LOCATION_PRINT(File, Loc) \
 
  40           location_print (File, Loc)
 
  42 static void version_check (location const *loc, char const *version);
 
  44 /* Request detailed syntax error messages, and pass them to GRAM_ERROR.
 
  45    FIXME: depends on the undocumented availability of YYLLOC.  */
 
  47 #define yyerror(Msg) \
 
  48         gram_error (&yylloc, Msg)
 
  49 static void gram_error (location const *, char const *);
 
  51 static char const *char_name (char);
 
  53 /** Add a lex-param or a parse-param.
 
  55  * \param type  \a lex_param or \a parse_param
 
  56  * \param decl  the formal argument
 
  57  * \param loc   the location in the source.
 
  59 static void add_param (char const *type, char *decl, location loc);
 
  62 static symbol_class current_class = unknown_sym;
 
  63 static uniqstr current_type = NULL;
 
  64 static symbol *current_lhs;
 
  65 static location current_lhs_location;
 
  66 static int current_prec = 0;
 
  68 #define YYTYPE_INT16 int_fast16_t
 
  69 #define YYTYPE_INT8 int_fast8_t
 
  70 #define YYTYPE_UINT16 uint_fast16_t
 
  71 #define YYTYPE_UINT8 uint_fast8_t
 
  85   /* Bison's grammar can initial empty locations, hence a default
 
  86      location is needed. */
 
  87   boundary_set (&@$.start, current_file, 1, 1);
 
  88   boundary_set (&@$.end, current_file, 1, 1);
 
 100   unsigned char character;
 
 103 /* Define the tokens together with their human representation.  */
 
 104 %token GRAM_EOF 0 "end of file"
 
 105 %token STRING     "string"
 
 108 %token PERCENT_TOKEN       "%token"
 
 109 %token PERCENT_NTERM       "%nterm"
 
 111 %token PERCENT_TYPE        "%type"
 
 112 %token PERCENT_DESTRUCTOR  "%destructor"
 
 113 %token PERCENT_PRINTER     "%printer"
 
 115 %token PERCENT_LEFT        "%left"
 
 116 %token PERCENT_RIGHT       "%right"
 
 117 %token PERCENT_NONASSOC    "%nonassoc"
 
 118 %token PERCENT_PRECEDENCE  "%precedence"
 
 120 %token PERCENT_PREC          "%prec"
 
 121 %token PERCENT_DPREC         "%dprec"
 
 122 %token PERCENT_MERGE         "%merge"
 
 125 /*----------------------.
 
 126 | Global Declarations.  |
 
 127 `----------------------*/
 
 131   PERCENT_DEFAULT_PREC    "%default-prec"
 
 132   PERCENT_DEFINE          "%define"
 
 133   PERCENT_DEFINES         "%defines"
 
 134   PERCENT_EXPECT          "%expect"
 
 135   PERCENT_EXPECT_RR       "%expect-rr"
 
 136   PERCENT_FLAG            "%<flag>"
 
 137   PERCENT_FILE_PREFIX     "%file-prefix"
 
 138   PERCENT_GLR_PARSER      "%glr-parser"
 
 139   PERCENT_INITIAL_ACTION  "%initial-action"
 
 140   PERCENT_LANGUAGE        "%language"
 
 141   PERCENT_LEX_PARAM       "%lex-param"
 
 142   PERCENT_NAME_PREFIX     "%name-prefix"
 
 143   PERCENT_NO_DEFAULT_PREC "%no-default-prec"
 
 144   PERCENT_NO_LINES        "%no-lines"
 
 145   PERCENT_NONDETERMINISTIC_PARSER
 
 146                           "%nondeterministic-parser"
 
 147   PERCENT_OUTPUT          "%output"
 
 148   PERCENT_PARSE_PARAM     "%parse-param"
 
 149   PERCENT_REQUIRE         "%require"
 
 150   PERCENT_SKELETON        "%skeleton"
 
 151   PERCENT_START           "%start"
 
 152   PERCENT_TOKEN_TABLE     "%token-table"
 
 153   PERCENT_VERBOSE         "%verbose"
 
 157 %token BRACED_CODE     "{...}"
 
 159 %token EPILOGUE        "epilogue"
 
 161 %token ID              "identifier"
 
 162 %token ID_COLON        "identifier:"
 
 163 %token PERCENT_PERCENT "%%"
 
 165 %token PROLOGUE        "%{...%}"
 
 171 %type <character> CHAR
 
 172 %printer { fputs (char_name ($$), stderr); } CHAR
 
 174 /* braceless is not to be used for rule or symbol actions, as it
 
 175    calls code_props_plain_init.  */
 
 176 %type <chars> STRING "%{...%}" EPILOGUE braceless content.opt
 
 178 %printer { fputs (quotearg_style (c_quoting_style, $$), stderr); }
 
 180 %printer { fprintf (stderr, "{\n%s\n}", $$); }
 
 181          braceless content.opt "{...}" "%{...%}" EPILOGUE
 
 183 %type <uniqstr> TAG ID ID_COLON PERCENT_FLAG variable
 
 184 %printer { fputs ($$, stderr); } ID variable
 
 185 %printer { fprintf (stderr, "%s:", $$); } ID_COLON
 
 186 %printer { fprintf (stderr, "%%%s", $$); } PERCENT_FLAG
 
 187 %printer { fprintf (stderr, "<%s>", $$); } TAG
 
 190 %printer { fprintf (stderr, "%d", $$); } INT
 
 192 %type <symbol> id id_colon symbol symbol.prec string_as_id
 
 193 %printer { fprintf (stderr, "%s", $$->tag); } id symbol string_as_id
 
 194 %printer { fprintf (stderr, "%s:", $$->tag); } id_colon
 
 196 %type <assoc> precedence_declarator
 
 197 %type <list>  symbols.1 symbols.prec generic_symlist generic_symlist_item
 
 201   prologue_declarations "%%" grammar epilogue.opt
 
 205         /*------------------------------------.
 
 206         | Declarations: before the first %%.  |
 
 207         `------------------------------------*/
 
 209 prologue_declarations:
 
 211 | prologue_declarations prologue_declaration
 
 214 prologue_declaration:
 
 218       code_props plain_code;
 
 219       code_props_plain_init (&plain_code, $1, @1);
 
 220       code_props_translate_code (&plain_code);
 
 221       gram_scanner_last_string_free ();
 
 222       muscle_code_grow (union_seen ? "post_prologue" : "pre_prologue",
 
 223                         plain_code.code, @1);
 
 224       code_scanner_last_string_free ();
 
 228       muscle_percent_define_ensure ($1, @1, true);
 
 230 | "%define" variable content.opt
 
 232       muscle_percent_define_insert ($2, @2, $3,
 
 233                                     MUSCLE_PERCENT_DEFINE_GRAMMAR_FILE);
 
 235 | "%defines"                       { defines_flag = true; }
 
 239       spec_defines_file = xstrdup ($2);
 
 241 | "%expect" INT                    { expected_sr_conflicts = $2; }
 
 242 | "%expect-rr" INT                 { expected_rr_conflicts = $2; }
 
 243 | "%file-prefix" STRING            { spec_file_prefix = $2; }
 
 244 | "%file-prefix" "=" STRING        { spec_file_prefix = $3; } /* deprecated */
 
 247       nondeterministic_parser = true;
 
 250 | "%initial-action" "{...}"
 
 253       code_props_symbol_action_init (&action, $2, @2);
 
 254       code_props_translate_code (&action);
 
 255       gram_scanner_last_string_free ();
 
 256       muscle_code_grow ("initial_action", action.code, @2);
 
 257       code_scanner_last_string_free ();
 
 259 | "%language" STRING            { language_argmatch ($2, grammar_prio, @1); }
 
 260 | "%lex-param" "{...}"          { add_param ("lex_param", $2, @2); }
 
 261 | "%name-prefix" STRING         { spec_name_prefix = $2; }
 
 262 | "%name-prefix" "=" STRING     { spec_name_prefix = $3; } /* deprecated */
 
 263 | "%no-lines"                   { no_lines_flag = true; }
 
 264 | "%nondeterministic-parser"    { nondeterministic_parser = true; }
 
 265 | "%output" STRING              { spec_outfile = $2; }
 
 266 | "%output" "=" STRING          { spec_outfile = $3; }  /* deprecated */
 
 267 | "%parse-param" "{...}"        { add_param ("parse_param", $2, @2); }
 
 268 | "%require" STRING             { version_check (&@2, $2); }
 
 271       char const *skeleton_user = $2;
 
 272       if (strchr (skeleton_user, '/'))
 
 274           size_t dir_length = strlen (current_file);
 
 275           char *skeleton_build;
 
 276           while (dir_length && current_file[dir_length - 1] != '/')
 
 278           while (dir_length && current_file[dir_length - 1] == '/')
 
 281             xmalloc (dir_length + 1 + strlen (skeleton_user) + 1);
 
 284               strncpy (skeleton_build, current_file, dir_length);
 
 285               skeleton_build[dir_length++] = '/';
 
 287           strcpy (skeleton_build + dir_length, skeleton_user);
 
 288           skeleton_user = uniqstr_new (skeleton_build);
 
 289           free (skeleton_build);
 
 291       skeleton_arg (skeleton_user, grammar_prio, @1);
 
 293 | "%token-table"                { token_table_flag = true; }
 
 294 | "%verbose"                    { report_flag |= report_states; }
 
 295 | "%yacc"                       { yacc_flag = true; }
 
 296 | /*FIXME: Err?  What is this horror doing here? */ ";"
 
 300   precedence_declaration
 
 304       grammar_start_symbol_set ($2, @2);
 
 306 | "%destructor" "{...}" generic_symlist
 
 309       for (list = $3; list; list = list->next)
 
 310         symbol_list_destructor_set (list, $2, @2);
 
 311       symbol_list_free ($3);
 
 313 | "%printer" "{...}" generic_symlist
 
 316       for (list = $3; list; list = list->next)
 
 317         symbol_list_printer_set (list, $2, @2);
 
 318       symbol_list_free ($3);
 
 326       default_prec = false;
 
 330       /* Do not invoke muscle_percent_code_grow here since it invokes
 
 331          muscle_user_name_list_grow.  */
 
 332       muscle_code_grow ("percent_code()", $2, @2);
 
 333       code_scanner_last_string_free ();
 
 335 | "%code" ID braceless
 
 337       muscle_percent_code_grow ($2, @2, $3, @3);
 
 338       code_scanner_last_string_free ();
 
 347 %token PERCENT_UNION "%union";
 
 351 | ID             { muscle_code_grow ("union_name", $1, @1); }
 
 355   "%union" union_name braceless
 
 358       muscle_code_grow ("stype", $3, @3);
 
 359       code_scanner_last_string_free ();
 
 367   "%nterm" { current_class = nterm_sym; } symbol_defs.1
 
 369       current_class = unknown_sym;
 
 372 | "%token" { current_class = token_sym; } symbol_defs.1
 
 374       current_class = unknown_sym;
 
 377 | "%type" TAG symbols.1
 
 381       for (list = $3; list; list = list->next)
 
 382         symbol_type_set (list->content.sym, $2, @2);
 
 383       symbol_list_free ($3);
 
 387 precedence_declaration:
 
 388   precedence_declarator tag.opt symbols.prec
 
 392       for (list = $3; list; list = list->next)
 
 394           symbol_type_set (list->content.sym, current_type, @2);
 
 395           symbol_precedence_set (list->content.sym, current_prec, $1, @1);
 
 397       symbol_list_free ($3);
 
 402 precedence_declarator:
 
 403   "%left"       { $$ = left_assoc; }
 
 404 | "%right"      { $$ = right_assoc; }
 
 405 | "%nonassoc"   { $$ = non_assoc; }
 
 406 | "%precedence" { $$ = precedence_assoc; }
 
 410   /* Nothing. */ { current_type = NULL; }
 
 411 | TAG            { current_type = $1; tag_seen = true; }
 
 414 /* Just like symbols.1 but accept INT for the sake of POSIX.  */
 
 417     { $$ = symbol_list_sym_new ($1, @1); }
 
 418 | symbols.prec symbol.prec
 
 419     { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
 
 424   | symbol INT { $$ = $1; symbol_user_token_number_set ($1, $2, @2); }
 
 427 /* One or more symbols to be %typed. */
 
 430     { $$ = symbol_list_sym_new ($1, @1); }
 
 432     { $$ = symbol_list_prepend ($1, symbol_list_sym_new ($2, @2)); }
 
 436   generic_symlist_item { $$ = $1; }
 
 437 | generic_symlist generic_symlist_item { $$ = symbol_list_prepend ($1, $2); }
 
 440 generic_symlist_item:
 
 441   symbol    { $$ = symbol_list_sym_new ($1, @1); }
 
 442 | TAG       { $$ = symbol_list_type_new ($1, @1); }
 
 443 | "<*>"     { $$ = symbol_list_default_tagged_new (@1); }
 
 444 | "<>"      { $$ = symbol_list_default_tagless_new (@1); }
 
 447 /* One token definition.  */
 
 456        symbol_class_set ($1, current_class, @1, true);
 
 457        symbol_type_set ($1, current_type, @1);
 
 461       symbol_class_set ($1, current_class, @1, true);
 
 462       symbol_type_set ($1, current_type, @1);
 
 463       symbol_user_token_number_set ($1, $2, @2);
 
 467       symbol_class_set ($1, current_class, @1, true);
 
 468       symbol_type_set ($1, current_type, @1);
 
 469       symbol_make_alias ($1, $2, @$);
 
 471 | id INT string_as_id
 
 473       symbol_class_set ($1, current_class, @1, true);
 
 474       symbol_type_set ($1, current_type, @1);
 
 475       symbol_user_token_number_set ($1, $2, @2);
 
 476       symbol_make_alias ($1, $3, @$);
 
 480 /* One or more symbol definitions. */
 
 483 | symbol_defs.1 symbol_def
 
 487         /*------------------------------------------.
 
 488         | The grammar section: between the two %%.  |
 
 489         `------------------------------------------*/
 
 492   rules_or_grammar_declaration
 
 493 | grammar rules_or_grammar_declaration
 
 496 /* As a Bison extension, one can use the grammar declarations in the
 
 497    body of the grammar.  */
 
 498 rules_or_grammar_declaration:
 
 500 | grammar_declaration ";"
 
 508   id_colon { current_lhs = $1; current_lhs_location = @1; } rhses.1
 
 512   rhs                { grammar_current_rule_end (@1); }
 
 513 | rhses.1 "|" rhs    { grammar_current_rule_end (@3); }
 
 519     { grammar_current_rule_begin (current_lhs, current_lhs_location); }
 
 521     { grammar_current_rule_symbol_append ($2, @2); }
 
 523     { grammar_current_rule_action_append ($2, @2); }
 
 525     { grammar_current_rule_prec_set ($3, @3); }
 
 527     { grammar_current_rule_dprec_set ($3, @3); }
 
 529     { grammar_current_rule_merge_set ($3, @3); }
 
 533 /*---------------------------.
 
 534 | variable and content.opt.  |
 
 535 `---------------------------*/
 
 537 /* The STRING form of variable is deprecated and is not M4-friendly.
 
 538    For example, M4 fails for `%define "[" "value"'.  */
 
 541 | STRING { $$ = uniqstr_new ($1); }
 
 544 /* Some content or empty by default. */
 
 546   /* Nothing. */   { $$ = ""; }
 
 558       code_props plain_code;
 
 559       $1[strlen ($1) - 1] = '\n';
 
 560       code_props_plain_init (&plain_code, $1+1, @1);
 
 561       code_props_translate_code (&plain_code);
 
 562       gram_scanner_last_string_free ();
 
 563       $$ = plain_code.code;
 
 572 /* Identifiers are returned as uniqstr values by the scanner.
 
 573    Depending on their use, we may need to make them genuine symbols.  */
 
 577     { $$ = symbol_from_uniqstr ($1, @1); }
 
 580       $$ = symbol_get (char_name ($1), @1);
 
 581       symbol_class_set ($$, token_sym, @1, false);
 
 582       symbol_user_token_number_set ($$, $1, @1);
 
 587   ID_COLON { $$ = symbol_from_uniqstr ($1, @1); }
 
 596 /* A string used as an ID: quote it.  */
 
 600       $$ = symbol_get (quotearg_style (c_quoting_style, $1), @1);
 
 601       symbol_class_set ($$, token_sym, @1, false);
 
 609       code_props plain_code;
 
 610       code_props_plain_init (&plain_code, $2, @2);
 
 611       code_props_translate_code (&plain_code);
 
 612       gram_scanner_last_string_free ();
 
 613       muscle_code_grow ("epilogue", plain_code.code, @2);
 
 614       code_scanner_last_string_free ();
 
 621 /* Return the location of the left-hand side of a rule whose
 
 622    right-hand side is RHS[1] ... RHS[N].  Ignore empty nonterminals in
 
 623    the right-hand side, and return an empty location equal to the end
 
 624    boundary of RHS[0] if the right-hand side is empty.  */
 
 627 lloc_default (YYLTYPE const *rhs, int n)
 
 632   /* SGI MIPSpro 7.4.1m miscompiles "loc.start = loc.end = rhs[n].end;".
 
 633      The bug is fixed in 7.4.2m, but play it safe for now.  */
 
 634   loc.start = rhs[n].end;
 
 635   loc.end = rhs[n].end;
 
 637   /* Ignore empty nonterminals the start of the the right-hand side.
 
 638      Do not bother to ignore them at the end of the right-hand side,
 
 639      since empty nonterminals have the same end as their predecessors.  */
 
 640   for (i = 1; i <= n; i++)
 
 641     if (! equal_boundaries (rhs[i].start, rhs[i].end))
 
 643         loc.start = rhs[i].start;
 
 651 /* Add a lex-param or a parse-param (depending on TYPE) with
 
 652    declaration DECL and location LOC.  */
 
 655 add_param (char const *type, char *decl, location loc)
 
 657   static char const alphanum[26 + 26 + 1 + 10] =
 
 658     "abcdefghijklmnopqrstuvwxyz"
 
 659     "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
 
 662   char const *name_start = NULL;
 
 665   /* Stop on last actual character.  */
 
 666   for (p = decl; p[1]; p++)
 
 668          || ! memchr (alphanum, p[-1], sizeof alphanum))
 
 669         && memchr (alphanum, p[0], sizeof alphanum - 10))
 
 672   /* Strip the surrounding '{' and '}', and any blanks just inside
 
 674   while (*--p == ' ' || *p == '\t')
 
 677   while (*++decl == ' ' || *decl == '\t')
 
 681     complain_at (loc, _("missing identifier in parameter declaration"));
 
 688            memchr (alphanum, name_start[name_len], sizeof alphanum);
 
 692       name = xmalloc (name_len + 1);
 
 693       memcpy (name, name_start, name_len);
 
 694       name[name_len] = '\0';
 
 695       muscle_pair_list_grow (type, decl, name);
 
 699   gram_scanner_last_string_free ();
 
 704 version_check (location const *loc, char const *version)
 
 706   if (strverscmp (version, PACKAGE_VERSION) > 0)
 
 708       complain_at (*loc, "require bison %s, but have %s",
 
 709                    version, PACKAGE_VERSION);
 
 715 gram_error (location const *loc, char const *msg)
 
 717   complain_at (*loc, "%s", msg);
 
 721 token_name (int type)
 
 723   return yytname[YYTRANSLATE (type)];
 
 734       buf[0] = '\''; buf[1] = c; buf[2] = '\''; buf[3] = '\0';
 
 735       return quotearg_style (escape_quoting_style, buf);